Introduction
Atrial fibrillation (AF) is the most common sustained cardiac arrhythmia in adults worldwide, and it is expected to increase, mainly due to the extended longevity in the overall population as well as the increasing burden of other comorbidities such as hypertension, diabetes, or heart failure (HF) [1–3]. Despite anticoagulation markedly decreasing the risk of stroke, patients remain at risk of cardiovascular disease, including coronary artery disease, HF and cardiovascular death [4, 5].
Heart failure and AF are two common conditions that frequently coexist. In fact, the presence of one entity may precipitate/exacerbate the other [5–7]. Remarkably, the increased risk of AF occurs in both, HF with reduced left ventricular ejection fraction (LVEF) (HFrEF) and HF with preserved LVEF (HFpEF) [8]. This is not surprising, as AF and HF share common risk factors and comorbidities. In addition, dilatation of left atrium, left atrial and ventricular fibrosis, chronic inflammation, neurohormonal hyperactivation, and electrophysiologic remodeling also play a relevant role [9, 10]. The concomitance of both conditions translates into higher morbidity and mortality rates, including a greater risk of thromboembolic events, and consequently, anticoagulation is recommended [6, 11–13]. Although a number of studies have analyzed the impact of HF on patients with AF taking vitamin K antagonists, the information currently available among patients treated with direct oral anticoagulants, particularly in clinical practice remains scarce [14–18].
In ROCKET-AF, rivaroxaban was noninferior to warfarin for the prevention of stroke and systemic embolic events and significantly reduced intracranial hemorrhage in patients with AF at high thromboembolic risk [19]. In a specific analysis of the ROCKET-AF trial, the relative efficacy and safety of rivaroxaban versus warfarin was independent of HF status [20]. However, data about the role of rivaroxaban among patients with HF and AF in clinical practice are warranted [21].
The EMIR (Estudio observacional para la identificación de los factores de riesgo asociados a eventos cardiovasculares mayores en pacientes con fibrilación auricular no valvular tratados con un anticoagulante oral directo [Rivaroxaban] [“Observational study to identify risk factors associated with major cardiovascular events in patients with nonvalvular atrial fibrillation treated with a direct oral anticoagulant [rivaroxaban]”) study [22, 23] was aimed to evaluate the performance of the cardiovascular risk 2MACE score in AF patients treated with rivaroxaban. In this study, the impact of the presence of HF at baseline on the clinical profile and outcomes in AF patients anticoagulated with rivaroxaban was analyzed.
Methods
The design and methods of the EMIR study have been extensively described in previous publications [22, 23]. Briefly, EMIR was a non-interventional and observational study that included patients 18 years or older, with an established diagnosis of AF (either paroxysmal, persistent or permanent), anticoagulated with rivaroxaban according to clinical practice ≥ 6 months before being enrolled and they provided written informed consent. Patients were recruited from 79 Spanish centers (hospitals and private clinics). By contrast, patients with prosthetic heart valves, any severe valvopathy, severe cognitive impairment, chronic infections or systemic autoimmune diseases, active cancer or severe liver insufficiency were excluded from the study. The study was approved by each participating Institutional Review Board.
Patients were followed-up during 2.5 years with 4 visits (baseline, 12 months, 24 months, and study end) that should coincide with any of the patients’ routine visits for HF management. No additional visits, laboratory tests, other diagnostic tests or treatments were specifically performed or prescribed for being included in the EMIR study. All data were recorded using an electronic case report form specifically created for the EMIR study.
At baseline, biodemographic data (age, sex, permanent AF, body mass index), risk stratification (CHA2DS2-VASc, HAS-BLED, and 2MACE score), cardiovascular risk factors (hypertension, diabetes), vascular disease (previous coronary artery disease, prior cerebrovascular disease, peripheral artery disease) and renal insufficiency were recorded. Data were collected from the clinical history of the patients and during the interview with the patient during the patients’ routine visit. The presence of HF was considered when it was reflected in the clinical history of the patient. Data were analyzed according to the presence of previous HF and the HF subtypes. HFrEF was defined as HF with a LVEF < 40%, HF with mildly reduced LVEF (HFmrEF) as HF with LVEF 40 – < 50% and HFpEF as HF with LVEF ≥ 50%. Renal insufficiency was defined as an estimated glomerular filtration rate < 60 mL/min/1.73 m2 by the MDRD-4 formula.
Events (major adverse cardiovascular events [MACE], thromboembolic events, myocardial infarction [MI], revascularization, cardiovascular death, death from any cause and major bleeding) during the study were evaluated. MACE events were defined as a combination of non-fatal MI, revascularization and cardiovascular death (death for coronary events, progressive HF death and sudden cardiac death). Thromboembolic events included stroke, systemic embolism and transient ischemic attack. Major bleedings were defined following the International Society of Thrombosis and Hemostasis definition [24]. The information source was in all cases the medical record and the patient. The investigator collected the study data from medical records or from personal interviews performed during the study follow-up. Before the present study started at the sites, all investigators were sufficiently trained on the background and objectives of the study. All outcome variables and covariates were recorded in a standardized electronic case report form. Medical review of the data was performed according to the medical review plan. A scientific committee independently evaluated and classified the events. Events were analyzed according to HF status and the HF subtypes. In addition, predictors of MACE, ischemic stroke and major bleeding in the EMIR population were analyzed.
Statement of ethics
This study protocol was reviewed and approved firstly by CAEIG (Comité Autonómico de Etica de Galicia) on July 14th, 2016, approval number 2016/348.
Patients were recruited from 79 Spanish centers (hospitals and private clinics). The study was approved by each participating Institutional Review Board.
Patients provided written informed consent.
Statistical methods
Qualitative variables were presented as absolute and relative frequencies and quantitative variables were described with measures of central tendency (mean and median) and dispersion (standard deviation and interquartile range). Qualitative variables were compared using the c2 test or the Fisher exact test, as required. When 2 means were compared, the t test or the Mann-Whitney test was used, when appropriate and 3 means (HF subtypes) by the Kruskal-Wallis test. Annual event rates were calculated. To assess predictors of MACE, thromboembolic events and major bleeding (dependent variables), multivariate analyzes were performed. The multivariate models began to be constructed by introducing those factors with a significance of p < 0.15 in the bivariates by the automatic variable selection method by steps forward and backward. Only the significant factors (p < 0.05) were finally considered to build the model. Odd ratios (OR) along with the 95% confidence interval (CI) were calculated. The following independent variables were considered: age (continuous variable), sex (female vs. male), body mass index (continuous variable), previous bleeding, diabetes, permanent AF, ischemic heart disease, coronary revascularization, antiplatelet agents, previous cerebrovascular disease, dependence level (dependent vs. autonomous), hypertension, hyperlipidemia, smoking, pulmonary disease, renal insufficiency, liver dysfunction, cancer, peripheral artery disease, alcohol use, non-severe dementia, HF, CHA2DS2-VASc (continuous variable), HAS-BLED (continuous variable) and 2MACE ≥ 3. A level of statistical significance of 0.05 was applied in all the statistical tests. The data were analyzed using the statistical package SPSS (v18.0 or superior).
Results
A total of 1,503 patients were initially enrolled. After the exclusion of 70 patients, 1,433 (93.7%) patients from 79 Spanish centers, were included for the final analysis, of whom 326 (22.7%) had HF at baseline (Fig. 1).
The baseline clinical characteristics of the overall study population and according to HF status are presented in Table 1A. Overall, mean age was 74.2 ± 9.7 years, 55.5% of patients were men, 37.5% had permanent AF, mean CHA2DS2-VASc score was 3.5 ± 1.5, mean HAS-BLED was 1.6 ± 1.0 and 26.9% had a 2MACE score ≥ 3. In addition, 79.3% of patients had hypertension, 27.1% diabetes, 28.3% vascular disease, 24.7% renal insufficiency (defined by MDRD-4 < 60 mL/ /min/1.73 m2), 16.4% ischemic heart disease and 12.5% prior cerebrovascular disease.
Among patients with HF (n = 326), 94 had HFrEF, 59 HFmrEF and 173 HFpEF. In patients with HF, mean LVEF was 48.0 ± 14.3%. Baseline clinical characteristics were analyzed according to HF status (Table 1A). Compared to patients without HF, patients with HF were older (75.3 ± 9.9 vs. 73.8 ± 9.6 years; p = 0.01), had more diabetes (36.5% vs. 24.3%; p < 0.01), permanent AF (50.9% vs. 33.3%; p < 0.01), previous coronary artery disease (28.2% vs. 12.9%; p < 0.01), renal insufficiency (MDRD-4: 31.7% vs. 22.6%; p = 0.001), as well as higher CHA2DS2-VASc score (4.5 ± 1.6 vs. 3.2 ± 1.4; p < 0.01), HAS-BLED score (1.8 ± ± 1.1 vs. 1.5 ± 1.0; p < 0.01) and more patients had a 2MACE score ≥ 3 (46.0% vs. 21.2%; p < 0.01). With regard to the baseline clinical characteristics according to HF subtype, patients with HFpEF were older, more commonly women, had a higher CHA2DS2-VASc score and more hypertension. By contrast, patients with HFrEF were more commonly men, had a higher 2MACE score ≥ 3 and more previous MI (Table 1B).
|
Total population (n = 1,433; 100%) |
HF population (n = 326; 22.7%) |
No HF population (n = 1,107;77.3%) |
P |
|
|
Biodemographic data |
||||
|
Age [years] |
74.2 ± 9.7 |
75.3 ± 9.9 |
73.8 ± 9.6 |
0.01 |
|
≥ 75 years |
691 (48.2%) |
173 (53.1%) |
518 (46.8%) |
0.05 |
|
Sex (men) |
795 (55.5%) |
193 (59.2%) |
602 (54.4%) |
0.12 |
|
Permanent atrial fibrillation |
535 (37.5%) |
166 (50.9%) |
369 (33.3%) |
< 0.01 |
|
Body mass index [kg/m2] |
29.1 ± 4.9 |
29.8 ± 5.3 |
28.9 ± 4.8 |
0.03 |
|
Risk stratification |
||||
|
CHA2DS2-VASc score |
3.5 ± 1.5 |
4.5 ± 1.6 |
3.2 ± 1.4 |
< 0.01 |
|
2MACE score ≥ 3 |
385 (26.9%) |
150 (46.0%) |
235 (21.2%) |
< 0.01 |
|
HAS-BLED score |
1.6 ± 1.0 |
1.8 ± 1.1 |
1.5 ± 1.0 |
< 0.01 |
|
Cardiovascular risk factors |
||||
|
Hypertension |
1,137 (79.3%) |
261 (80.1%) |
876 (79.1%) |
0.72 |
|
Diabetes |
388 (27.1%) |
119 (36.5%) |
269 (24.3%) |
< 0.01 |
|
Vascular disease |
||||
|
Vascular disease |
406 (28.3%) |
127 (39.0%) |
279 (25.2%) |
< 0.01 |
|
Previous coronary disease |
235 (16.4%) |
92 (28.2%) |
143 (12.9%) |
< 0.01 |
|
Prior cerebrovascular disease |
179 (12.5%) |
42 (12.9%) |
137(12.4%) |
0.81 |
|
Peripheral artery disease and/or aortic plaque |
96 (6.7%) |
33 (10.1%) |
63 (5.7%) |
0.005 |
|
Peripheral artery disease |
58 (4.0%) |
22 (6.7%) |
36 (3.3%) |
0.005 |
|
Other conditions/comorbidities |
||||
|
Renal insufficiency (MDRD-4: < 60 mL/min/1.73 m2) |
350 (24.7%) |
103 (31.7%) |
247 (22.6%) |
0.01 |
|
Renal insufficiency (Cockcroft-Gault: < 60 mL/min/1.73 m2) |
498 (35.1%) |
133 (40.8%) |
365 (33.0%) |
0.01 |
|
HFpEF |
HFmrEF |
HFrEF |
P |
|
|
Biodemographic data |
||||
|
N |
173 |
59 |
94 |
|
|
Proportion in the overall population |
12.1% |
4.1% |
6.6% |
– |
|
Proportion in the HF population |
53.1% |
18.1% |
28.8% |
|
|
Age [years] |
77.5 ± 9.3 |
73.3 ± 9.8 |
72.6 ± 10.2 |
< 0.001 |
|
≥ 75 years |
111 (64.2%) |
23 (39.0%) |
39 (41.5%) |
< 0.001 |
|
Sex (men) |
79 (45.7%) |
42 (71.2%) |
72 (76.6%) |
< 0.001 |
|
Permanent atrial fibrillation |
79 (45.7%) |
30 (50.8%) |
51 (54.3%) |
0.39 |
|
Body mass index [kg/m2] |
30.6 ± 5.5 |
29.0 ± 5.3 |
28.8 ± 4.7 |
0.016 |
|
Risk stratification |
||||
|
CHA2DS2-VASc score |
4.9 ± 1.4 |
4.0 ± 1.5 |
4.2 ± 1.8 |
< 0.001 |
|
2MACE score ≥ 3 |
71 (41.0%) |
21 (35.6%) |
58 (61.7%) |
0.001 |
|
HAS-BLED score |
1.9±1.0 |
1.7±1.1 |
1.6±1.2 |
0.15 |
|
Cardiovascular risk factors |
||||
|
Arterial hypertension |
149 (86.1%) |
45 (76.3%) |
67 (71.3%) |
0.011 |
|
Diabetes |
65 (37.6%) |
21 (35.6%) |
33 (35.1%) |
0.91 |
|
Vascular disease |
||||
|
Vascular disease |
59 (34.1%) |
24 (40.7%) |
44 (46.8%) |
0.12 |
|
Previous coronary disease |
39 (22.5%) |
20 (33.9%) |
33 (35.1%) |
0.053 |
|
Previous myocardial infarction |
12 (6.9%) |
10 (16.9%) |
25 (26.6%) |
< 0.001 |
|
Previous cerebrovascular disease |
23 (13.3%) |
5 (8.5%) |
14 (14.9%) |
0.50 |
|
Peripheral artery disease and/or aortic plaque |
17 (9.8%) |
8 (13.6%) |
8 (8.5%) |
0.59 |
|
Peripheral artery disease |
9 (5.2%) |
7 (11.9%) |
6 (6.4%) |
0.21 |
|
Other conditions/comorbidities |
||||
|
Renal insufficiency (MDRD-4: < 60 mL/min/1.73 m2) |
51 (29.5%) |
19 (32.2%) |
33 (35.5%) |
0.60 |
|
Renal insufficiency (Cockcroft-Gault: < 60 mL/min/1.73 m2) |
70 (40.5%) |
24 (40.7%) |
39 (41.9%) |
0.97 |
The mean follow-up was 2.2 ± 0.6 years (median 2.5 years, interquartile range 2.2–2.6 years). The annual rates of relevant events were calculated over 1,425 patients (323 out of 326 patients with HF and 1,102 out of 1,107 patients without HF) (Table 2A). Overall, 87 (6.1%) patients died during the study period, of whom 20 (1.4%) had a cardiovascular origin, where 13 (0.9%) were due to progressive chronic HF. As a result, 70.0% of cardiovascular deaths were caused by progressive HF. Among patients with baseline HF, annual rates of (stroke + systemic embolism + transient ischemic attack), MACE, cardiovascular death, death from any cause and major bleeding were 1.2%, 3.0%, 2.0%, 5.5% and 1.4%, respectively. Compared to those patients without HF at baseline, those patients with HF had greater annual rates of MACE (3.0% vs. 0.5%; p < 0.01), cardiovascular death (2.0% vs. 0.2%; p < 0.01) and death from any cause (5.5% vs. 2.0%; p < 0.01), without significant differences regarding other outcomes, including thromboembolic or bleeding events. With regard to events during the follow-up according to HF subtype, patients with HFpEF had more thromboembolic events and patients with HFrEF more MACE, MI and cardiovascular death (Table 2B).
|
Total |
HF (n = 326) |
No HF (n = 1,107) |
P |
|
|
Stroke + SE + TIA: |
||||
|
Number patients (%) |
23 (1.6) |
8 (2.5) |
15 (1.4) |
0.17 |
|
Annual rate events (%)* |
0.7 |
1.2 |
0.6 |
0.22 |
|
Major bleeding: |
||||
|
Number patients (%) |
29 (2.0) |
8 (2.5) |
21 (1.9) |
0.53 |
|
Annual rate events (%)* |
1.0 |
1.4 |
0.9 |
0.33 |
|
MACE: |
||||
|
Number patients (%) |
30 (2.1) |
17 (5.2) |
13 (1.2) |
< 0.01 |
|
Annual rate events (%)* |
1.1 |
3.0 |
0.5 |
< 0.01 |
|
Myocardial infarction: |
||||
|
Number patients (%) |
5 (0.3) |
3 (0.9) |
2 (0.2) |
0.08 |
|
Annual rate events (%)* |
0.2 |
0.4 |
0.1 |
0.15 |
|
Revascularization: |
||||
|
Number patients (%) |
9 (0.6) |
4 (1.2) |
5 (0.5) |
0.13 |
|
Annual rate events (%)* |
0.3 |
0.6 |
0.2 |
0.22 |
|
Cardiovascular death: |
||||
|
Number patients (%) |
20 (1.4) |
13 (4.0) |
7 (0.6) |
< 0.01 |
|
Annual rate events (%)* |
0.6 |
2.0 |
0.2 |
< 0.01 |
|
Death from any cause: |
||||
|
Number patients (%) |
87 (6.1) |
38 (11.7) |
49 (4.4) |
< 0.01 |
|
Annual rate events (%)* |
2.7 |
5.5 |
2.0 |
< 0.01 |
|
Type of heart failure |
P |
|||
|
HFpEF (n = 173) Annual rate of events (n = 171; accumulated time = 386.81 years) |
HFmrEF (n = 59) Annual rate of events (n = 58; accumulated time = 128.29 years) |
HFrEF (n = 94) Annual rate of events (n = 94; accumulated time = 179.80 years) |
||
|
Stroke + SE + TIA: |
||||
|
Number patients (%) |
8 (4.6) |
0 |
0 |
0.036 |
|
Annual rate events (%)* |
2.07 |
0 |
0 |
HFpEF vs. HFrEF: p < 0.05 |
|
Major bleeding: |
||||
|
Number patients (%) |
6 (3.5) |
0 |
2 (2.1) |
0.392 |
|
Annual rate events (%)* |
1.81 |
0 |
1.67 |
No statistically significant difference |
|
MACE: |
< 0.001 |
|||
|
Number patients (%) |
2 (1.2) |
1 (1.7) |
14 (14.9) |
HFpEF vs. HFrEF: p < 0.001 |
|
Annual rate events (%)* |
0.78 |
0.78 |
9.45 |
HFmrEF vs. HFrEF: p < 0.001 |
|
Myocardial infarction: |
||||
|
Number patients (%) |
0 |
0 |
3 (3.2) |
0.029 |
|
Annual rate events (%)* |
0 |
0 |
1.67 |
HFpEF vs. HFrEF: p < 0.05 |
|
Revascularization: |
||||
|
Number patients (%) |
1 (0.6) |
0 |
3 (3.2) |
0.110 |
|
Annual rate events (%)* |
0.26 |
0 |
1.67 |
No statistically significant difference |
|
Cardiovascular death: |
< 0.001 |
|||
|
Number patients (%) |
2 (1.2) |
1 (1.7) |
11 (11.7) |
HFpEF vs. HFrEF: p < 0.001 |
|
Annual rate events (%)* |
0.52 |
0.78 |
6.12 |
HFmrEF vs. HFrEF: p < 0.05 |
|
Death from any cause: |
||||
|
Number patients (%) |
14 (8.1) |
6 (10.2) |
18 (19.1) |
0.025 |
|
Annual rate events (%)* |
3.62 |
4.68 |
10.01 |
HFpEF vs. HFrEF: p < 0.05 |
A multivariate logistic regression analysis was performed to study the potential predictors of MACE events, thromboembolic events and major bleeding (Table 3). The presence of ischemic heart disease, renal insufficiency and HF were independent predictive factors associated to MACE in the global population. Type of AF did not have an impact on MACE risk (p = 0.662). On the other hand, the use of antiplatelet agents, non-severe dementia and CHA2DS2-VASc score were independently associated with the development of thromboembolic events and a score 2MACE ≥ 3, dependency and HAS-BLED score with major bleeding. The main results of the study are presented in Figure 2.
|
Independent variables |
Univariate analyzis |
Multivariate analyzis |
||||
|
P |
OR |
95% CI |
P |
OR |
95% CI |
|
|
Dependent variable “MACE events” |
||||||
|
Antiplatelet agents |
< 0.01 |
13.6 |
6.3–29.6 |
|||
|
Heart failure |
< 0.01 |
4.7 |
2.2–10.1 |
0.002 |
3.4 |
1.6–7.3 |
|
Coronary revascularization |
< 0.01 |
4.6 |
2.1–10.1 |
|||
|
Ischemic heart disease |
< 0.01 |
4.6 |
2.2–9.8 |
0.002 |
3.4 |
1.6–7.3 |
|
2MACE ≥ 3 |
0.002 |
3.2 |
1.5–6.9 |
|||
|
Renal insufficiency |
0.007 |
3.0 |
1.4–6.5 |
0.02 |
2.5 |
1.2–5.5 |
|
Peripheral artery disease |
0.08 |
2.9 |
0.9–10.1 |
|||
|
Diabetes |
0.15 |
1.8 |
0.8–3.8 |
|||
|
HAS-BLED (continuous variable) |
0.02 |
1.5 |
1.1–2.1 |
|||
|
CHA2DS2-VASC (continuous variable) |
0.05 |
1.3 |
1.0–1.6 |
|||
|
Sex (female vs. male) |
0.02 |
0.3 |
0.1–0.8 |
|||
|
Dependent variable “Thromboembolic events” |
||||||
|
Antiplatelet agents |
< 0.01 |
9.2 |
3.7–22.7 |
< 0.01 |
9.0 |
3.5–23.0 |
|
Non-severe dementia |
0.002 |
7.7 |
2.2–27.5 |
0.02 |
5.5 |
1.3–22.7 |
|
2MACE ≥ 3 |
< 0.01 |
4.5 |
1.9–11.1 |
|||
|
Heart failure |
0.002 |
3.8 |
1.6–9.1 |
|||
|
Previous bleeding |
0.12 |
3.3 |
0.7–14.5 |
|||
|
Diabetes |
0.01 |
3.0 |
1.3–7.2 |
|||
|
Previous stroke |
0.03 |
2.8 |
1.1–7.4 |
|||
|
Coronary revascularization |
0.04 |
2.8 |
1.1–7.3 |
|||
|
Ischemic heart disease |
0.04 |
2.6 |
1.0–6.5 |
|||
|
CHA2DS2-VASc (continuous variable) |
0.002 |
1.5 |
1.2–1.9 |
0.01 |
1.4 |
1.1–1.8 |
|
Age (continuous variable) |
0.15 |
1.0 |
1.0–1.1 |
|||
|
Dependent variable “Major bleeding” |
||||||
|
Arterial hypertension |
0.049 |
7.5 |
1.0–55.0 |
|||
|
Non-severe dementia |
0.009 |
5.3 |
1.5–18.4 |
|||
|
Previous bleeding |
0.003 |
5.2 |
1.7–15.6 |
|||
|
Patient autonomy (dependent vs. autonomous) |
< 0.01 |
4.7 |
2.1–10.7 |
0.03 |
2.6 |
1.1–6.1 |
|
2MACE ≥ 3 |
< 0.01 |
4.6 |
2.2–9.9 |
0.02 |
2.6 |
1.1–6.1 |
|
Renal insufficiency |
< 0.01 |
4.4 |
2.1–9.3 |
|||
|
Previous stroke |
0.004 |
3.2 |
1.4–7.2 |
|||
|
Antiplatelet agents |
0.03 |
3.0 |
1.1–8.0 |
|||
|
HAS-BLED (continuous variable) |
< 0.01 |
2.2 |
1.6–3.0 |
0.01 |
1.9 |
1.3–2.7 |
|
Coronary revascularization |
0.07 |
2.2 |
0.9–5.3 |
|||
|
Diabetes |
0.08 |
1.9 |
0.9–4.1 |
|||
|
Hyperlipidemia |
0.14 |
1.8 |
0.8–4.1 |
|||
|
CHA2DS2-VASc (continuous variable) |
< 0.01 |
1.6 |
1.3–2.0 |
|||
|
Age (continuous variable) |
< 0.01 |
1.1 |
1.0–1.1 |
|||
Discussion
The present study showed that in a wide sample of real-life patients with AF anticoagulated with rivaroxaban compared to those patients without HF at baseline, individuals with previous HF have a worse clinical profile and a higher risk of MACE (cardiac mortality, coronary revascularization, non-fatal MI) and cardiovascular mortality. In addition, the history of HF is independently associated with the development of MACE, but not with stroke or major bleeding. Despite that, rates of MACE and death remained low in HF patients, indicating that anticoagulation with rivaroxaban may be a good choice in this population. This information is relevant, as although previous studies have analyzed the impact of HF on outcomes in anticoagulated AF patients, very scarce information is available in those patients taken rivaroxaban [21].
In the EMIR study, nearly 23% of AF patients presented with HF at baseline. This is in line with previous studies performed in Spain that have shown that the concomitance of both conditions is very common in clinical practice. Thus, in the PAULA study, that included AF patients treated in a primary care setting and anticoagulated with vitamin K antagonists, approximately 24% of patients had HF [25]. More recently, in the FANTASIIA study that included AF patients in a specialized cardiology setting and anticoagulated with direct oral anticoagulants or vitamin K antagonists, nearly 30% of patients also had HF [26]. In the international GLORIA-AF registry, in which AF patients anticoagulated with a direct oral anticoagulant were enrolled, 24% of patients had HF at baseline [27]. Conversely, different studies have shown that approximately one third of patients with HF also have AF [28, 29]. As each entity enhances the development of the other one [5, 6], an active search should be promoted to rule out the concomitance of both conditions [1].
In the current study, more patients had HFpEF than HFrEF or HFmrEF. Although some authors have not shown significant differences in the strength of an association between AF and the type of HF [8], other authors have reported that among HF patients, AF is progressively more common with increasing LVEF [28, 30, 31]. Despite previous studies showing that HFpEF accounts for at least half of the cases of HF, it is very likely that due to the ageing of the population, this proportion will increase in the following years, as well as the number of patients with AF and HF concomitantly [32, 33]. Remarkably, the present study showed that there were relevant differences in the clinical profile of patients according to HF subtype, particularly related with age, sex and some comorbidities. These differences are in line with previous studies of HF population [34, 35].
Compared to patients without HF, patients with HF had a worse clinical profile, with more risk factors, and comorbidities, as well as a greater thromboembolic and bleeding risk. This high-risk profile in patients with HF has also been observed in previous studies [29]. As a result, to reduce the disease burden in this population, all patients with HF and AF should receive in addition to anticoagulation, guideline-adherent HF therapy [36].
Remarkably, different studies have shown that the superiority of direct oral anticoagulants over vitamin K antagonists remain in patients with HF, in both, clinical trials and real-life studies [16–18]. In addition, it is more difficult to attain an adequate time in therapeutic range among patients taking vitamin K antagonists in patients with HF, leading to a lower protection [15]. In the current study, all patients were taking rivaroxaban. After a median follow-up of 2.5 years, among patients with previous HF, annual rates of thromboembolic events, cardiovascular death, death from any cause and major bleeding were 1.2%, 2.0%, 5.5%, and 1.4%, respectively. In the rivaroxaban arm of the ROCKET-AF trial, these numbers were 1.9%, 3.4%, 5.1%, and 14.2% for major or nonmajor clinically relevant bleeding, respectively [20]. A retrospective study performed in the United States that analyzed patients with HF and AF taking rivaroxaban between 2011 and 2016 showed that after a median follow-up of 1.4 years, rates of stroke or systemic embolism, ischemic stroke, major bleeding and intracranial hemorrhage were 1.0, 0.7, 3.9, and 0.3 events per 100 person-years, respectively [21]. As a result, in clinical practice, rates of outcomes in patients with HF and AF seem lower than those reported in the ROCKET-AF trial [20].
Although in non-anticoagulated patients the most devastating consequence related to AF is stroke and its associated complications (death and disability), anticoagulation changes mortality and outcome patterns in AF patients. Thus, in the anticoagulated AF population, most deaths are cardiac-related (cardiovascular death, MI and HF), and only a small proportion are associated with stroke and bleeding [37]. This has also been described in the AF population with HF, including those patients with HF enrolled in the ROCKET-AF trial [18, 20, 38]. Likewise, in the present study, compared to those patients without HF at baseline, those patients with HF had greater annual rates of MACE and cardiovascular death, but with similar rates of thromboembolic and bleeding events. In addition, the multivariate analyzes showed that previous HF was independently associated with the development of MACE, but not with thromboembolic events or major bleeding. Therefore, anticoagulation is not only important in AF patients with HF, but also choosing an oral anticoagulant that effectively reduces MACE events [1]. In this context, experimental and clinical studies have shown that rivaroxaban decreases the progression of ischemic cardiomyopathy, as well as the risk of MI and cardiovascular death [39–42]. Of note, the use of antiplatelet agents was independently associated with the development of thromboembolic events. In the AFIRE trial, rivaroxaban monotherapy was noninferior to a combination of rivaroxaban with an antiplatelet agent for thromboembolic events or death, and superior for major bleeding in AF patients with stable coronary artery disease, and this occurred irrespective of their risk for stroke and bleeding [43]. Therefore, all these data indicate that rivaroxaban in monotherapy should be considered in patients with HF and AF, not only to reduce the risk of thromboembolic events, but also the risk of MACE and cardiovascular mortality, leading to a comprehensive management of this population. On the other hand, despite the AMADEUS study, which showed that the risk of cardiovascular death, stroke, or systemic embolism increased among patients with permanent AF (vs. nonpermanent AF), regardless the presence of HF [44], in the present study the type of AF was not associated with an increased risk of outcomes. However, it should be noted that idraparinux and vitamin K antagonists were the anticoagulants used in AMADEUS, compared with rivaroxaban in the study herein. On the other hand, the current study showed that the risk of events varied according to HF subtype (thromboembolic events in HFpEF and MACE, MI and cardiovascular death in HFrEF). Other studies have also shown differences in outcomes according to HF subtype [45]. As a result, these particularities should be taken into account to provide a comprehensive approach in the management of patients with AF and HF.
Limitations of the study
As this was an observational study, no control group was available, and the presence of some confounding factors could not be excluded. However, the high number of patients included, as well as that the recruitment was performed consecutively after office consultation, may reduce possible selection bias. On the other hand, it should be considered that the patients were recruited after at least 6 months under rivaroxaban treatment. Therefore, the results of this study can only be extended to a similar population.
Conclusions
Nearly 1 out of 4 patients with AF anticoagulated with rivaroxaban in clinical practice have HF concomitantly. After a median follow-up of 2.5 years, annual rates of thromboembolic events, MACE, cardiovascular death, and major bleeding in HF population are 1.2%, 3.0%, 2.0%, and 1.4%, respectively. Compared to patients without HF, HF patients are older, have a greater baseline risk profile, and a higher risk of developing MACE and cardiovascular mortality, but not thromboembolic or bleeding events. The management of patients with HF and AF requires a comprehensive approach, with the aim to reduce not only the stroke risk, but also cardiovascular-related complications. In this context, rivaroxaban should be considered as a first-line therapy in the treatment of patients with HF and AF in clinical practice.
Acknowledgments
Writing and editorial assistance was provided by Content Ed Net (Madrid, Spain) with funding from Bayer Hispania.
Funding
The EMIR Study was funded by Bayer Hispania SL.
